Patients with Parkinson's Disease have exceptionally poor sleep. Even relative to other neurodegenerative diseases such as Alzheimer's Disease, the sleep of the PD patient is fragmented and disturbed. Most notably, sleep in PD is characterized by excessive activity in surface electomyographic (EMG) recordings from many different muscle groups. Despite the sleep disturbance, approximately 50 percent of PD patients note that, on nights when they are able to achieve sleep, they experience a transient (1 to 3 hour) reduction in waking motor symptoms upon arising in the morning. This effect has been termed Sleep Benefit. To date there are no double-blind placebo-controlled studies treating sleep disturbance using PD using any (conventional or alternative) medical treatments. In this randomized, double- blind, parallel-groups, placebo-controlled polysomnographic clinical trial we will compare two alternative medical treatments (valerian, melatonin) and two conventional medical treatments (diphenhydramine, zolpidem) for the disturbed sleep of PD patients. Compelling basic science and clinical rationales exist for use of each of these substances (including valerian and melatonin) for treatment of sleep disturbance in PD. The proposed study will be conducted for six consecutive nights (3 Baseline, 3 Drug) using state-of the-art digitized ambulatory polysomnography in each patient's home. Outcomes will include both measures of nocturnal sleep and waking motor function. Polysomnographic measurements will include customary variables such as total sleep time, sleep efficiency and sleep latency, as well as EMG measures of periodic and isolated muscle activity during sleep. Assessments of motor function will be made the morning immediately, following the third Baseline night and third Drug night in order to test for improvement related to improved sleep. Data will be analyzed with Analysis of Covariance examining Condition (Baseline, Drug), (valerian, melatonin, diphenhydramine, zolpidem, placebo), and characteristic Sleep Benefit (positive, negative) main effects, as well as their interactions, after adjusting for Baseline values. The results would represent the first data applying rigorous clinical trial methodology to the study of disturbance sleep in PD patients and would critically examine the efficacy of two substances currently seeing widespread use as over-the counter hypnotics for which little polysomnographic data currently exist.